Overview of Tabulation and Analysis in MR

While there are no standardized tabulation and analysis procedure that should always be followed, it is quite common for researchers to proceed through the following six steps

Preparing the raw data: Each completed data-collection form must be edited to assure that the data contained therein are legible and accurate. Then, for a number of questions on the questionnaire, it will be necessary to establish information categories and classes in agreement with the items specified on the list of needed information. For example, in the yogurt study, it may be desirable to classify each household’s yogurt consumption into only three or four categories (e.g. heavy consumers, moderate consumers, light consumers) rather than have many categories ranging from one container to perhaps as many as two dozen containers. After these categories are established each data collection form is reviewed to identify the consumption of each household and to classify it into one of three or four selected classes of consumption.

Overview of tabulation and Analysis:
1. Preparing the raw data: editing and coding
2. Entering the data into the computer
3. Tabulating the data
4. Determining whether significant differences exist between categories
5. Explaining why differences exist
6. Making recommendations

Entering Data into the computer: After the data collective forms have all been edited and the responses all placed into the proper classes and categories the data must be entered into the computer.

Tabulating the Data:
The Computer can be programmed to count the number of responses falling into each response category of any question and it can even make counts of responses to two or three questions simultaneously. Summary measures such as percentages, ranges, and frequency distributions can also be determined by the computer during this counting activity.

Determining whether significant differences exist between categories: Typically researchers will observe differences among the data in different categories (e.g. households in rural areas eat less yogurt than households in urban areas) Researchers will determine these differences are too large to have occurred by chance as a result of sampling variations and, if so, will conclude that they reflect true differences between the categories. Statistical test of significance are used of this purpose.

Explaining “Why” Differences Exist: If it is determined that the differences observed between categories are statistically significant, an attempt should be made to explain why there are significant differences. Researchers who do not attempt to explain the differences may be overlooking important findings. This, in turn, may cause them to draw unwarranted conclusions.

This “why” information my require that the observed differences be further analyzed using still other information obtained in the study. For example analysis may show the rural area households have lower incomes than do urban area households and this may help explain why rural area households consume less yogurt than urban area households.

Making Recommendations:
After drawing statistical conclusions, the analyst needs to translate them into recommendations. Making recommendations usually requires an understanding of the practical details surrounding a given operation and so may not be the responsibility of researchers. In general, however, when researchers are qualified by their general knowledge of the operations, they should recommend.

It should be made clear that the tabulation and analysis function does not always follow precisely the six step procedure outlined above. In many projects the steps will tend to overlap, since at any time the analyst may generate new hypotheses that require recycling earlier steps. All of the steps however, are involved to some degree in each analysis.

The procedures presented in the remainder of this article show how researchers can perform the first three tabulation steps in a manner which should result in a minimum of errors. The three methods of testing for the statistical significance of observed differences most commonly used in marketing research.

The following discussions on data preparation and tabulation apply to all types of data collection forms – questionnaires, panel diaries, and forms used in observation studies. All such forms are similar in that they record some kind of response. To simplify the presentation, reference will only be made to questionnaires, even though the discussion applies equally to all data collection forms.